Roger Mayer RM57 build?

[RuudNL]

OK, here is the first version of the prototype. Some component values may change.

The modifications: the value of the series resistor from the input to the FET is now 18 K.
This reduces the noise significantly.

Also C3/R4/R12 reduce distortion, by modulating the gate of the FET with the audio signal.

P7 should be adjusted with the wiper of P3 to ground, in such a way that the audio signal is just not attenuated.
The correct value of the voltage over C9 depends on the FET used.  It may be necessary to change the value of R18 and/or P7.

P3 sets the level at which the signal is limited.

 

[Spencerleehorton]

Great work!!! I would be interested in hearing this, might knock up a version, if that’s ok ruud?

 

[RuudNL]

Feel free to build one!
It may be necessary to optimize some component values, but the result is pretty good already.
Keep in mind that the TL074 needs symmetrical power, so +15V and - 15V on pins 4 and 11.
I have calibrated the prototype in such a way that I get 1 V AC on the 'top' of P2. This means 30 mV AC over the FET.
Signal to noise ratio ('A' weighted) is in the order of 80 dB re. 1 V AC, so very acceptable.
Next thing I would like to optimize is distortion at low frequencies with fast release times. (Always a problem.)
I have already changed a couple of component values, and added a 'composite release' switch (optional).
With this switch closed, you get a fast release on isolated peaks but a longer release on more constant limiting.

 

[Spencerleehorton]

Ok my problems so far are mains hum on one channel and one channel is missing a release pot.
I must say though it’s a very musical  compression, I will try and connect up transformers to get it balanced and get both channels working the same then look at why I’m getting main hum.
Maybe I don’t have Psu connected correctly?
Or maybe I need better filtered psu?
My earth is connected to chassis but no where else, pretty sure this is correct.

 

[ilfungo]

Where could be the meter connected?

 

[RuudNL]

In case you are referring to the '2018' version: you can use the same part of the original schematic to connect a gain reduction meter.

 

[Rob Flinn]

Ok my problems so far are mains hum on one channel and one channel is missing a release pot.
I must say though it’s a very musical  compression, I will try and connect up transformers to get it balanced and get both channels working the same then look at why I’m getting main hum.
Maybe I don’t have Psu connected correctly?
Or maybe I need better filtered psu?
My earth is connected to chassis but no where else, pretty sure this is correct.

You may be getting an issue because the unit is unbalanced.  I would try just touching a wire between 0v & Chassis ground and see if it goes, 10 sec work & you know the answer.......

 

[Spencerleehorton]

Ok rob, good shout, will try in a bit.

 

[RuudNL]

Here is a picture of an original RM57 PCB. Just in case someone wants an example of the layout.

 

[Spencerleehorton]

Question, i am getting quite a lot of compression, is there a way to reduce the amount of compression?
with attack CCW and release at fastest, bringing up the threshold does kick the compressor in quite hard, i wondered if there is a way of reducing this.

Also are we able to figure out where all the trimmers should be set or at least have some sort of guess?

Mains hum still present so far, will try another PSU and see if it cleans it up, so far having chassis connected to pin 1 gives mains hum. Haven't checked with this disconnected yet.

 

[RuudNL]

The input of the limiter is pretty sensitive, -30 dB or so.
This is done intentionally, to reduce distortion, because the (AC) voltage over the FET should be low.
There are two ways to reduce the 'threshold':

1- Reduce the level before the input and add extra gain after the limiter to restore the level.
Signal to noise ratio will suffer.

2- Reduce the level going to the detector. (The signal going to the base of Q9. Increase the value of VR5)
Disadvantage is that the distortion will be higher.

 

[Spencerleehorton]

any idea what sort of signal to noise the original is compared to your version?

Your version sounds very promising, its a very musical compressor, i've tried it on drums and bass and so far its very fun to use, just a shame (the old version) has its limitations.

 

[RuudNL]

I would like to improve the metering.
A meter range of 40 dB is ridiculous. (If someone wants to limit 40 dB, he is crazy anyway!)
Personally I think a meter range of 20 dB is more than adequate.
Another problem is that 5 dB gain reduction is about half the scale...
What I would like to see, is the use of a standard VU meter, that can be used to measure the output level OR the gain reduction.
In such a way that the meter indicates 0 dB without any gain reduction going on, and the needle moving backwards when gain reduction occurs.
It seems this is the way how it is done in the newer issues of the RM58. (The mono box.)

And about the signal to noise ratio: in my "2018" version this is close to 80 dB re. 1 Volt. ("A" weighted.)
This value is also mentioned in the specifications of the original, but I doubt very much if this is true...

By the way: interesting to see what the going price of this compressor is nowadays: €7,714.85 for a handful components...
https://reverb.com/item/11866606-crazy-rare-roger-mayer-rm-57-stereo-compressor-from-the-record-plant-in-nyc-modded-bra  ;D

 

[Vac11]

Thank you Ruud, for your work and great info. Pleas could you recommend us input and output transformer for balancing  input and output of original circuit? Which ratio? Resistance of primary and secondary?

 

[RuudNL]

If you would like to add an input transformer, I suppose a 1:1 input transformer with a specified impedance of 10 K would be suitable.
Because the input is pretty sensitive, it might be a good idea to put an attenuator in front of the transformer.
This way the signal level for the transformer will be lower.
(For example 3.3K in series with each input wire, and a third 3.3K resistor in parallel with the primary.
This will produce about 10 dB of attenuation and lowers the LF distortion in the transformer.)

For the output you should think about a 1:1 transformer with a specified impedance of 600 ohms.
Because signal levels are pretty high here, the transformer should have a core that is not too small.
A small modem transformer will work, but you will lose the lowest frequencies.
Also the core will saturate at low frequencies.
Most standard audio line level output transformers will work fine.

 

[Spencerleehorton]

Would you be able to put up a drawing of how the input and output should be with transformers please.
I always find it much less confusing.

Could I use OEP 2;1 txf set to 1:1 and use resistors?
Or should I wind my own?
I have made several 10468 txf in mumetal cases with 80% nickel core which may help?

 

[Spencerleehorton]

I tried wiring up the oep txf and the output level was very low and much more distortion, so I’ve either wired it up wrong or they are not the best type for this application?

 

[RuudNL]

Connection of an input/output transformer can be as simple as this.
Small transformers won't work properly as output transformer.
For input you will need a (relative) high-Z transformer, for the output a low-Z one.
Usually you will connect the balanced input- and output wires to pins 2 and 3 of and XLR connector. Pin 1 connects to ground.
The R/C at the output is optional and depends on the transformer used. (To optimize frequency/square wave response.)
Keep in mind that the output capability of a TL074 is limited, don't expect +26 dBm!

 

[Rob Flinn]

Would a 5534 be better on the output ?  At least it has some drive capability and in conjunction with a complementary pair of power transistors is quite good

 

[Spencerleehorton]

i'm a sucker for a drawing!!! thanks ruud

would a 10468 type transformer be ok to use?